English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The challenges inherent to long-horizon tasks in robotics persist due to thetypical inefficient exploration and sparse rewards in traditional reinforcementlearning approaches. To alleviate these challenges, we introduce a novelalgorithm, Variational Autoencoder-based Subgoal Inference (VAESI), toaccomplish long-horizon tasks through a divide-and-conquer manner. VAESIconsists of three components: a Variational Autoencoder (VAE)-based SubgoalGenerator, a Hindsight Sampler, and a Value Selector. The VAE-based SubgoalGenerator draws inspiration from the human capacity to infer subgoals andreason about the final goal in the context of these subgoals. It is composed ofan explicit encoder model, engineered to generate subgoals, and an implicitdecoder model, designed to enhance the quality of the generated subgoals bypredicting the final goal. Additionally, the Hindsight Sampler selects validsubgoals from an offline dataset to enhance the feasibility of the generatedsubgoals. The Value Selector utilizes the value function in reinforcementlearning to filter the optimal subgoals from subgoal candidates. To validateour method, we conduct several long-horizon tasks in both simulation and thereal world, including one locomotion task and three manipulation tasks. Theobtained quantitative and qualitative data indicate that our approach achievespromising performance compared to other baseline methods. These experimentalresults can be seen in the website\url{https://sites.google.com/view/vaesi/home}.

Implicit Subgoal Planning with Variational Autoencoders for Long-Horizon Sparse Reward Robotic Tasks